We propose to conduct a five-year prospective longitudinal study of language development in adolescents with fragile X syndrome (FXS). The data collected will address three specific aims. (1) We will describe the developmental trajectory of important, theoretically motivated domains of language and the ways in which those trajectories differ for males and females with FXS. In doing so, we will use nonverbal cognitive development as a benchmark and thereby determine whether language poses a challenge beyond that created by general cognitive limitations. (2) We will identify the determinants of within-syndrome variation in language development and differences in those determinants across males and females with FXS. In doing so, we will focus on the adolescent's psychological and behavioral characteristics (e.g., memory, autism status), his or her biological characteristics (e.g, FMRP levels), and the supportiveness of the environment, particularly the mother. Moderating and mediating relationships involving selected predictors will also be tested. (3) We will begin the process of distinguishing between those properties of the language development profile of FXS that are specific to it rather than shared with other forms of mental retardation. In doing so, we will compare the developmental trajectories of language in FXS and Down syndrome (DS). DS is a useful comparison because it overlaps substantially with FXS in terms of overall severity of impairment and because its phenotype contrasts with that of FXS in ways that can illuminate the factors affecting language development. Data will be collected at yearly intervals from adolescent boys and girls with FXS and from younger typically developing (TD) children and adolescents with DS. The TD children and adolescents with DS will be selected so that over the course of the project they traverse the same period of nonverbal cognitive development as do the adolescents with FXS. In addition, the comparison between FXS and DS will involve groups matched on age and nonverbal IQ. Measures will include standardized tests, experimental tasks, and informant reports. Dependent variables will be composites of variables reflecting important conceptual distinctions in understanding language development (e.g., mastery of forms vs. mastery of the social uses of language, expression vs. reception). Hierarchical Linear Modeling (HLM) will be used to analyze the data.